Treadmill exercise promotes retinal astrocyte plasticity and protects against retinal degeneration in a mouse model of light-induced retinal degeneration

Abstract

Exercise is an effective neuroprotective intervention that preserves retinal function and structure in several animal models of retinal degeneration. However, the retinal cell types governing exercise-induced neuroprotection remain elusive. Previously, we found exercise-induced retinal neuroprotection was associated with increased levels of retinal brain-derived neurotrophic factor (BDNF) and required intact signal transduction with its high-affinity receptor, tropomyosin kinase B (TrkB). Brain studies have shown astrocytes express BDNF and TrkB and that decreased BDNF-TrkB signaling in astrocytes contributes to neurodegeneration. Additionally, exercise has been shown to alter astrocyte morphology. Using a light-induced retinal degeneration (LIRD) model, we investigated how exercise influences retinal astrocytes in adult male BALB/c mice. Treadmill exercise in dim control and LIRD groups had increased astrocyte density, GFAP labeling, branching, dendritic endpoints, and arborization. Meanwhile, inactive LIRD animals had significant reductions in all measured parameters. Additionally, exercised groups had increased astrocytic BDNF expression that was visualized using proximity ligase assay. Isolated retinal astrocytes from exercised LIRD groups had significantly increased expression of a specific isoform of TrkB associated with cell survival, TrkB.FL. Conversely, inactive LIRD isolated retinal astrocytes had significantly increased expression of TrkB.T1, which has been implicated in neuronal cell death. Our data indicate exercise not only alters retinal astrocyte morphology but also promotes specific BDNF-TrkB signaling associated with cell survival and protection during retinal degeneration. These findings provide novel insights into the effects of treadmill exercise on retinal astrocyte morphology and cellular expression, highlighting retinal astrocytes as a potential cell type involved in BDNF-TrkB signaling.

Keywords: RRID:AB_2039891; RRID:AB_2534102; RRID:AB_2894998; RRID:AB_788229; RRID:AB_880202; aerobic exercise; light damage model; retina; retinal neuroprotection.

FIGURE 1

Retinal astrocyte labeling and density are increased with exercise. Active groups were exercised by treadmill once a day for 2 weeks at a speed of 10 m/min for 60 min, meanwhile inactive groups were placed on static treadmills for the same duration (a). Light-induced retinal degeneration (LIRD) occurred during the second week of exercise using light exposure of 5000 lux for 4 h. Following retinal functional testing by electroretinography (ERG) in the third week, mice were euthanized and retinal tissue was collected for histology, immunofluorescence, and retinal astrocyte isolation. Retinal flat mounts from inactive + dim (b), active + dim (c), inactive + LIRD (d), and active + LIRD (e) mice were stained for glial fibrillary acidic protein (GFAP, green). Cells positively labeled for GFAP were quantified by fluorescence (f), GFAP expression (g), and number of astrocytes per image (h) and two-way ANOVA with Tukey’s multiple comparison analyses were performed. Retinal astrocytes in both active groups had increased GFAP fluorescence as well as an increased in the number of positively labeled astrocytes present. A.U., arbitrary units, *p = .0129, **p = .0052, ***p < .001, ****p < .0001, scale bar = 20 μm. Fluorescence and astrocyte density quantifications were performed using n = 6 animals per group, each symbol in the plots represents the average of three biological replicates per retinal quadrant. For GFAP expression, bar graph represents mean values with each data point representing n = 4, n = 12 total per group. The experimenter was blinded to experimental conditions and animal group IDs. Values are mean ± SD

FIGURE 2

Treadmill exercise increased astrocytic branching and endpoint voxels in the retina. Retinal astrocytes from each experimental group were analyzed using Skeletonize plugin from ImageJ (a–d). Astrocyte branches and endpoint voxels (dendritic endings) were then quantified (e and f, respectively) and two-way ANOVA with Tukey’s multiple comparison analyses were performed. Active + LIRD retinal astrocytes appeared to be morphologically similar to inactive + dim astrocytes and had no statistically significant differences in branching and endpoint number. In contrast, inactive + LIRD retinal astrocytes had significantly decreased branching and endpoint voxels. ****p < .0001, scale bar = 20 μm. N = 4 animals per group, five biological replicates = 20 cells per group. Experimenter was blinded to experimental conditions and animal group IDs. Values are mean ± SD

FIGURE 3

Treadmill running increased retinal astrocyte complexity. Retinal astrocytes from each experimental group were analyzed using Sholl analysis (a–d) and astrocyte dendritic complexity was quantified by counting the number of intersections from the soma (e). Retinal astrocytes from active + dim animals had a significant increase in dendritic complexity compared to inactive + dim animals (7–12 μm from the soma), whereas inactive + LIRD retinal astrocytes had a significant decrease in the number of intersections compared to inactive + dim animals (8–10 μm, 12–15 μm and 17 μm). Active + LIRD retinal astrocytes showed a decrease in branching at 13–17 μm compared to inactive + dim animals. Two-way ANOVA with Dunnett’s multiple comparisons test was performed. Significance between inactive + dim and each experimental group is indicated by the following symbols: active + dim = ^, inactive + LIRD = #, active + LIRD = †; *p < .05, **p < .01, ***p < .001, scale bar = 20 μm. N = 4 animals per group, three biological replicates = 12 cells per group. Experimenter was blinded to experimental conditions and animal group IDs. Values are mean ± SD

FIGURE 4

Increased retinal BDNF–astrocyte interaction observed in treadmill exercised mice. Proximity ligase assay (PLA) was performed on retinal flat mounts from inactive (a and c) and active (b and d) dim and LIRD exposed mice using antibodies targeting BDNF and glutamate transporter-1 (GLT1). GLT1 is an extracellular membrane bound transporter expressed in astrocytes and endothelial cells. Positive co-labeling for astrocytes was done using GFAP (red). Green fluorescence indicates BDNF-GLT1 interaction (expressed in endothelial cells), yellow fluorescence (overlap of green fluorescence from PLA and red fluorescence from GFAP labeling) signifies co-labeling of retinal astrocytes and BDNF. Quantification of PLA fluorescence (green fluorescence overlapping with red fluorescence) revealed a significant increase in BDNF-GLT1 interaction in retinal astrocytes from active groups compared to inactive groups (e). ****p < .0001, scale bars = 15 μm, white box signifies magnified region of image outlined in blue box. N = 4 animals per group, each symbol in the plot represents the average of three biological replicates per retinal quadrant. Experimenter was blinded to experimental conditions and animal group IDs. Two-way ANOVA with Tukey’s multiple comparison analysis was performed. Values are mean ± SD

FIGURE 5

Magnetic-activated cell sorting reveals exercise alters BDNF and specific TrkB isoform expression in isolated retinal astrocytes. Isolated retinal astrocytes from experimental groups were probed for BDNF (a) and its high-affinity receptor, TrkB, in its catalytically active (TrkB.FL, b) and truncated (TrkB.T1, c) isoforms. Inactive + LIRD retinal astrocytes showed a significant decrease in BDNF expression and a significant increase in TrkB.T1 expression, which has been associated with neuronal cell death. Active + LIRD retinal astrocytes have BDNF expression similar to dim treated groups and had a significant increase in TrkB.FL, which has been associated with cell survival. Bar graphs represent mean values with each data point representing n = 4, n = 12 total per group. Two-way ANOVA with Tukey’s multiple comparison analysis was performed. *p < .05, **p < .01, values are mean ± SD